Some text processing tasks are easier to read when using continuation and ... "pipe"

pipe-nltk-chunking.py

import pprint
from pipe import Pipe, select, as_list
import nltk

stokenize = Pipe(lambda text: nltk.sent_tokenize(text))
wtokenize = Pipe(lambda sentences: sentences | select(nltk.word_tokenize))
tag = Pipe(lambda sentences: sentences | select(nltk.pos_tag))
chunk = Pipe(lambda sentences: sentences | select(nltk.ne_chunk))

text = ("A Wiki is a website which is editable over the web by it's users. "
          "This allows information to be more rapidly updated than traditional websites.")

text | stokenize | wtokenize | tag | select(pprint.pprint) | as_list

'''The output should be something like this:

[('A', 'DT'),
 ('Wiki', 'NNP'),
 ('is', 'VBZ'),
 ('a', 'DT'),
 ('website', 'JJ'),
 ('which', 'WDT'),
 ('is', 'VBZ'),
 ('editable', 'JJ'),
 ('over', 'IN'),
 ('the', 'DT'),
 ('web', 'NN'),
 ('by', 'IN'),
 ('it', 'PRP'),
 ("'s", 'VBZ'),
 ('users', 'NNS'),
 ('.', '.')]
[('This', 'DT'),
 ('allows', 'VBZ'),
 ('information', 'NN'),
 ('to', 'TO'),
 ('be', 'VB'),
 ('more', 'JJR'),
 ('rapidly', 'RB'),
 ('updated', 'VBN'),
 ('than', 'IN'),
 ('traditional', 'JJ'),
 ('websites', 'NNS'),
 ('.', '.')]
[None, None]
'''